In [2]:

# -*- coding: utf-8 -*-
								"""
								Created on Viernes 8 Octubre 2021 17:43:11 
								
								@author: Daniel Busso
								"""
								
								# Plantilla de Pre Procesado
								
								# Cómo importar las librerías
								import numpy    as np
								import pandas   as pd
								import datetime as dt
								import os
								
								# Gráficos
								# ==============================================================================
								import matplotlib.pyplot as plt
								import seaborn as sns
								%matplotlib inline
								#plt.style.use('fivethirtyeight')
								
								#from statsmodels.graphics.tsaplots import plot_acf
								#from statsmodels.graphics.tsaplots import plot_pacf
								
								# Configuración warnings
								# ==============================================================================
								import warnings
								warnings.filterwarnings('ignore') 
								

In [3]:

# Modelado y Forecasting
								# ==============================================================================
								from sklearn.linear_model import LinearRegression
								from sklearn.linear_model import Lasso
								from sklearn.linear_model import Ridge
								from sklearn.ensemble import RandomForestRegressor
								#from xgboost import XGBRegressor
								
								from sklearn.metrics import mean_squared_error
								from sklearn.metrics import mean_absolute_error
								#from sklearn.metrics import mean_absolute_percentage_error
								
								#from skforecast.ForecasterAutoreg import ForecasterAutoreg
								#from skforecast.ForecasterAutoregMultiOutput import ForecasterAutoregMultiOutput
								#from skforecast.model_selection import grid_search_forecaster
								#from skforecast.model_selection import time_series_spliter
								#from skforecast.model_selection import cv_forecaster
								#from skforecast.model_selection import backtesting_forecaster
								#from skforecast.model_selection import backtesting_forecaster_intervals
								

In [4]:

# Configuracion Pandas
								
								pd.set_option('display.max_rows', 200)
								pd.set_option('display.max_columns', None)
								pd.set_option('display.width', None)
								pd.set_option('display.max_colwidth', -1)
								pd.set_option('display.float_format', '{:.2f}'.format)
								

In [5]:

def conversion_2400(fecha):
									if fecha[11:13] != '24':
										return pd.to_datetime(fecha, format='%d/%m/%Y %H:%M')
								
									fecha = fecha[0:11] + '00:' + fecha[14:]
									return pd.to_datetime(fecha, format='%d/%m/%Y %H:%M') + dt.timedelta(days=1)
								

In [6]:

# EMPIEZO PROCESANDO UNA CARPETA
								# AUTOMATIZAR LECTURAS DE VARIOS ARCHIVOS .PRN ANUALES
								# lEER Y CONCATENAR LOS ARCHIVOS .prn DE LA CARPETA DSPractica/eetts/sanfrancisco
								
								import os
								import sys
								from pathlib import Path
								
								# ELEGIR PLATAFORMA WINDOWS LOCAL O GOOGLE COLABORATORY (la Nube)
								if sys.platform == 'linux' :
								  from google.colab import drive
								  drive.mount('/content/drive', force_remount=True)
								  carpeta_trabajo = Path('/content/drive/Othercomputers/LENOVO/eett_prod/san_francisco/Demanda_13') # Google Colab
								else:
								  carpeta_inicio  = Path.home()
								  carpeta_eett = Path("eett_prod/san_francisco/Demanda_13")
								  carpeta_trabajo = carpeta_inicio / carpeta_eett
								
								archivos = list(carpeta_trabajo.glob('*.prn'))
								
								# Automatizacion de los DataFrames 
								dfconca = pd.DataFrame([])
								for nombre_completo in archivos:
									dfdatos = pd.read_fwf(nombre_completo, header=0, decimal=',')
									dfconca = pd.concat([dfconca, dfdatos], ignore_index=True)
									
								dfconca    
								

Out[6]:

	Fecha	Hora	MW	MVAR	MVA	Ampere	KVolt
0	01/01/2010	00:15	15.80	7.13	17.33	711.24	0.00
1	01/01/2010	00:30	15.63	6.82	17.05	706.06	0.00
2	01/01/2010	00:45	15.49	6.80	16.91	696.61	0.00
3	01/01/2010	01:00	15.55	6.81	16.98	698.32	0.00
4	01/01/2010	01:15	15.50	6.82	16.93	694.93	0.00
...	...	...	...	...	...	...	...
411454	31/12/2021	23:00	23.25	-2.69	23.41	980.38	4.46
411455	31/12/2021	23:15	22.86	-3.06	23.07	971.29	4.44
411456	31/12/2021	23:30	22.66	-3.20	22.89	972.08	4.43
411457	31/12/2021	23:45	22.77	7.82	24.08	993.98	4.45
411458	31/12/2021	24:00	22.49	7.86	23.82	981.86	4.45

411459 rows × 7 columns

In [7]:

archivos

Out[7]:

[WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2010_1201.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2011_1200.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2012_1159.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2013_1157.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2014_1156.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2015_1154.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2016_1152.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2017_1151.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2018_1150.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2019_1150.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2020_1148.prn'),
								 WindowsPath('C:/Users/daniel/eett_prod/san_francisco/Demanda_13/MED_DEM_SANFCO13_2021_1246.prn')]

In [8]:

#
								# BORRAMOS COLUMNAS NO USADAS
								#
								
								del dfconca['Ampere']
								del dfconca['KVolt']
								

In [9]:

# GENERAMOS INDICE DE FECHAS PARA MANEJAR LA SERIE DE TIEMPO
								
								dfconca['Fecha_Hora'] = dfconca['Fecha'] + ' ' + dfconca['Hora']
								dfconca['Fecha_Hora'] = dfconca.Fecha_Hora.apply(conversion_2400) # CONVERSION DE HORA 24:00 ---> 00:00
								
								dfconca.set_index('Fecha_Hora', inplace=True)
								
								serie     = dfconca.drop(dfconca.index[-1]) # borro la fila del nuevo anio
								serie     = serie.drop(columns=['Fecha', 'Hora'])
								
								serie_ord = serie.sort_index()
								
								datos = serie_ord.copy()
								demandas = datos.copy()
								demandas
								

Out[9]:

	MW	MVAR	MVA
Fecha_Hora
2010-01-01 00:15:00	15.80	7.13	17.33
2010-01-01 00:30:00	15.63	6.82	17.05
2010-01-01 00:45:00	15.49	6.80	16.91
2010-01-01 01:00:00	15.55	6.81	16.98
2010-01-01 01:15:00	15.50	6.82	16.93
...	...	...	...
2021-12-31 22:45:00	23.42	-2.79	23.58
2021-12-31 23:00:00	23.25	-2.69	23.41
2021-12-31 23:15:00	22.86	-3.06	23.07
2021-12-31 23:30:00	22.66	-3.20	22.89
2021-12-31 23:45:00	22.77	7.82	24.08

411458 rows × 3 columns

In [10]:

datos

Out[10]:

	MW	MVAR	MVA
Fecha_Hora
2010-01-01 00:15:00	15.80	7.13	17.33
2010-01-01 00:30:00	15.63	6.82	17.05
2010-01-01 00:45:00	15.49	6.80	16.91
2010-01-01 01:00:00	15.55	6.81	16.98
2010-01-01 01:15:00	15.50	6.82	16.93
...	...	...	...
2021-12-31 22:45:00	23.42	-2.79	23.58
2021-12-31 23:00:00	23.25	-2.69	23.41
2021-12-31 23:15:00	22.86	-3.06	23.07
2021-12-31 23:30:00	22.66	-3.20	22.89
2021-12-31 23:45:00	22.77	7.82	24.08

411458 rows × 3 columns

In [11]:

if datos.index.has_duplicates:
									print('SI TIENE DUPLICADOS')
								else:
									print('NO TIENE DUPLICADOS')
								

NO TIENE DUPLICADOS

In [12]:

#dti2 = pd.date_range(start='2017-01-01 00:15:00', end='2018-01-01 00:00:00', freq="15min")
								#
								# No se si es tan completo porque a partir del anio 2016 las mediciones van cada 5 min
								indice_completo = pd.date_range(start=datos.index[0], end=datos.index[-1], freq="15min")
								#
								dife = indice_completo.difference(datos.index)
								
								if dife.empty == False:
									print('Faltan algunas Mediciones: ', len(dife), '% Incompletas= ', (len(dife)/len(indice_completo)*100), '% Completas= ', \
										  (((len(indice_completo)-len(dife))/len(indice_completo))*100))
								else:
									print('\n')
									print('Mediciones Completas: ', (((len(indice_completo)-len(dife))/len(indice_completo))*100))
								

Faltan algunas Mediciones:  9309 % Incompletas=  2.2123883289326396 % Completas=  97.78761167106737

In [13]:

dife

Out[13]:

DatetimeIndex(['2010-08-14 00:00:00', '2010-11-10 00:15:00',
											   '2010-11-10 00:30:00', '2010-11-10 00:45:00',
											   '2010-11-10 01:00:00', '2010-11-10 01:15:00',
											   '2010-11-10 01:30:00', '2010-11-10 01:45:00',
											   '2010-11-10 02:00:00', '2010-11-10 02:15:00',
											   ...
											   '2021-09-17 08:45:00', '2021-09-17 09:00:00',
											   '2021-09-17 09:15:00', '2021-09-17 09:30:00',
											   '2021-09-18 08:45:00', '2021-09-18 09:00:00',
											   '2021-09-18 09:15:00', '2021-09-18 09:30:00',
											   '2021-12-03 09:00:00', '2021-12-03 09:15:00'],
											  dtype='datetime64[ns]', length=9309, freq=None)

In [ ]:

In [14]:

#
								# OBTENGO LAS MEDICIONES FALTANTES FECHAS-HORA SIN MEDICION
								
								dsm = pd.DataFrame({'Mediciones Faltantes':dife}, index=dife)
								dsm
								

Out[14]:

	Mediciones Faltantes
2010-08-14 00:00:00	2010-08-14 00:00:00
2010-11-10 00:15:00	2010-11-10 00:15:00
2010-11-10 00:30:00	2010-11-10 00:30:00
2010-11-10 00:45:00	2010-11-10 00:45:00
2010-11-10 01:00:00	2010-11-10 01:00:00
...	...
2021-09-18 09:00:00	2021-09-18 09:00:00
2021-09-18 09:15:00	2021-09-18 09:15:00
2021-09-18 09:30:00	2021-09-18 09:30:00
2021-12-03 09:00:00	2021-12-03 09:00:00
2021-12-03 09:15:00	2021-12-03 09:15:00

9309 rows × 1 columns

In [15]:

# AGRUPO LAS MEDICIONES FALTANTES
								#
								
								dsm['anio'] = dsm.index.year
								dsm['mes']  = dsm.index.month
								dsm['dia']  = dsm.index.day
								dsm_grp = dsm.groupby(by=['anio','mes', 'dia'])
								
								med_faltantes = dsm_grp.count()
								med_faltantes
								
								# MEDICIONES FALTANTES PARA UN ANIO EN PARTICULAR
								#med_faltantes.loc[(2010), :]
								
								# MEDICIONES FALTANTES PARA UN DIA EN PARTICULAR
								#
								#med_faltantes.loc[(2010,11,10), :]
								
								
								#dsm_grp.get_group((2010, 11, 10))
								#dsm_grp.groups
								#dsm_grp.indices
								

Out[15]:

			Mediciones Faltantes
anio	mes	dia
2010	8	14	1
	11	10	95
	11	11	1
2011	2	3	1
2012	1	12	95
...	...	...	...
2021	8	13	5
	8	27	1
	9	17	6
	9	18	4
	12	3	2

278 rows × 1 columns

In [17]:

dsm

Out[17]:

	Mediciones Faltantes	anio	mes	dia
2010-08-14 00:00:00	2010-08-14 00:00:00	2010	8	14
2010-11-10 00:15:00	2010-11-10 00:15:00	2010	11	10
2010-11-10 00:30:00	2010-11-10 00:30:00	2010	11	10
2010-11-10 00:45:00	2010-11-10 00:45:00	2010	11	10
2010-11-10 01:00:00	2010-11-10 01:00:00	2010	11	10
...	...	...	...	...
2021-09-18 09:00:00	2021-09-18 09:00:00	2021	9	18
2021-09-18 09:15:00	2021-09-18 09:15:00	2021	9	18
2021-09-18 09:30:00	2021-09-18 09:30:00	2021	9	18
2021-12-03 09:00:00	2021-12-03 09:00:00	2021	12	3
2021-12-03 09:15:00	2021-12-03 09:15:00	2021	12	3

9309 rows × 4 columns

In [22]:

med_faltantes.loc[(2014), :]
								

Out[22]:

		Mediciones Faltantes
mes	dia
3	19	6
5	16	1
6	19	95
	20	96
	21	96
	22	96
	23	1

In [ ]:

In [15]:

# Gráfico boxplot para estacionalidad anual 
								# Cada Box Plot grafica el mes para los 12 anios 
								# Por ejemplo Octubre '10' son todos los valores de Octubre en los 12 anios
								# ==============================================================================
								fig, ax = plt.subplots(figsize=(10, 5))
								plt.style.use('fivethirtyeight')
								
								datos['mes'] = datos.index.month
								datos.boxplot(column='MVA', by='mes', ax=ax)
								datos.groupby('mes')['MVA'].median().plot(style='o-', linewidth=0.8, ax=ax)
								
								# ESTETICA DEL GRAFICO
								plt.xticks(fontsize=12)  #plt.xticks(fontsize=15, rotation=90)
								plt.yticks(fontsize=12)
								#ax.set_ylabel('Demandas')
								ax.set_title('Distribución Demandas por mes 2010 - 2021')
								ax.legend()
								fig.suptitle('');
								

In [16]:

# Gráfico boxplot para estacionalidad anual 
								# Blox Plot para todos los anios
								# ==============================================================================
								
								def Grafica_Caja(pi, pf):
									
									plt.style.use('fivethirtyeight')
									fig = plt.figure()
									fig, ax = plt.subplots(figsize=(30, 5), dpi=100)  # (15, 5)
									
									# PREPARO LOS DATOS
									datos['mes']  = datos.index.month
									datos['anio'] = datos.index.year
									datos2 = datos[pi:pf]
									
									# GRAFICO EL BOXPLOT
									datos2.boxplot(column='MVA', by=['anio','mes'], ax=ax)
									
									# GRAFICO LA LINEA DE MEDIANA
									y = datos2.groupby(by=['anio','mes'])['MVA'].median()
									pos_sup = y.index.shape[0] + 1
									x = np.arange(1, pos_sup)
									ax.plot(x, y, lw=1.5, label='Mediana [MVA]', marker='o')
									#datos3.groupby(by=['anio','mes'])['MVA'].median().plot(style='o-', linewidth=0.8, ax=ax)
									
									# ESTETICA DEL GRAFICO
									plt.xticks(fontsize=12, rotation=90)  #plt.xticks(fontsize=15, rotation=90)
									plt.yticks(fontsize=12)
									#ax.set_ylabel('Demandas')
									ax.set_title('Distribución Demandas por Mes ' +pi+' - '+pf)
									ax.legend() 
									fig.suptitle('');
								

In [17]:

Grafica_Caja('2010', '2021')
								

<Figure size 432x288 with 0 Axes>

In [18]:

Grafica_Caja('2017', '2017')
								

<Figure size 432x288 with 0 Axes>

In [19]:

Grafica_Caja('2019', '2019')
								

<Figure size 432x288 with 0 Axes>

In [20]:

Grafica_Caja('2020', '2020')
								

<Figure size 432x288 with 0 Axes>

In [21]:

# CALCULO FORMAL DE LOS LIMITES SUPERIOR E INFERIOR
								# Demanda = Transformacion 132/13 Kv + Generacion Propia
								# Demanda = TD1 + TD2 + TD3 + TG1 + TG2
								
								#Pot_Ins_Tra = pd.Series([25.0, 25.0, 25.0], index=['TD1', 'TD2', 'TD3'])
								#POT_INS_TRAFOS = Pot_Ins_Tra.sum()
								
								#Pot_Ins_Gen = pd.Series([20.0, 28.0], index=['TG1', 'TG2'])  # con cos fi = 0.8 TG1 = 16 MW y TG2 = 23 MW 
								#POT_INS_GENE = Pot_Ins_Gen.sum()
								#POT_TOTAL = POT_INS_TRAFOS + POT_INS_GENE
								
								
								#indice_ini = datos.index[0] 
								#delta = pd.Timedelta(value=90, unit='D')
								#indice_fin = indice_ini + delta
								#verano = datos.loc[indice_ini:indice_fin, 'MVA'] 
								
								#===============================================================================================================
								# LA IDEA ES CALCULAR LOS LIMITES PARA PERIODOS ESTIVALES DE 6 MESES
								# PERIODO VERANO1  '2020-01-01':'2020-03-31'
								# PERIODO INVIERNO '2020-04-01:2020-09-30'
								# PERIODO VERANO2  '2020-10-01:2020-12-31'
								#
								def lim_sup(q75, q25):
									IQR = q75 - q25
									bigote = 3 * IQR       # LO NORMAL SE USA 1.5
									return (q75 + bigote)
									
								atipicos = pd.Series([])    
								anios = ['2010', '2011', '2012', '2013', '2014', '2015', '2016', '2017', '2018', '2019', '2020', '2021']
								for anio in anios:
									ver1i = anio + '-01-01 00' 
									ver1f = anio + '-03-31 23'
									invi  = anio + '-04-01 00' 
									invf  = anio + '-09-30 23'
									ver2i = anio + '-10-01 00' 
									ver2f = anio + '-12-31 23'
									
									#verano1  = demandas.loc[ver1i:ver1f, 'MVA'].resample('D').max()
									#invierno = demandas.loc[invi:invf, 'MVA'].resample('D').max()
									#verano2  = demandas.loc[ver2i:ver2f, 'MVA'].resample('D').max()
									verano1  = demandas.loc[ver1i:ver1f, 'MVA']
									invierno = demandas.loc[invi:invf, 'MVA']
									verano2  = demandas.loc[ver2i:ver2f, 'MVA']
									
									lim_sup_ver1 = lim_sup(verano1.quantile(0.75), verano1.quantile(0.25)) 
									lim_sup_inv  = lim_sup(invierno.quantile(0.75), invierno.quantile(0.25)) 
									lim_sup_ver2 = lim_sup(verano2.quantile(0.75), verano2.quantile(0.25)) 
									
									# REMOVER EJEMPLOS DE ATIPICOS 
									#indice1 = pd.to_datetime('2010-01-01 19:08:08')
									#indice  = pd.to_datetime('2010-04-01 19:12:12')
									#indice2 = pd.to_datetime('2010-10-01 19:14:14')
									#demandas.loc[indice1, 'MVA'] = 150.0
									#demandas.loc[indice, 'MVA'] = 150.0
									#demandas.loc[indice2, 'MVA'] = 150.0
									
									# GUARDAR LOS ATIPICOS A REMOVER Y LOS VALORES
									atipicos_ver1 = verano1[verano1   > lim_sup_ver1] 
									atipicos_inv  = invierno[invierno > lim_sup_inv]
									atipicos_ver2 = verano2[verano2   > lim_sup_ver2]    
									atipicos = pd.concat([atipicos, atipicos_ver1, atipicos_inv, atipicos_ver2]) 
								
								atipicos_ord = atipicos.sort_index()
								#datos = demandas.drop(atipicos.index)
								atipicos_ord
								

Out[21]:

2017-01-27 09:30:00   75.78
								2017-04-22 08:15:00   28.43
								2017-04-22 08:30:00   28.23
								2017-04-22 08:45:00   42.52
								2017-04-22 09:00:00   41.99
													   ... 
								2020-08-22 11:45:00   22.96
								2020-08-22 12:15:00   22.92
								2020-08-22 13:30:00   23.74
								2020-08-22 13:45:00   23.29
								2021-05-04 11:30:00   59.22
								Length: 476, dtype: float64

In [22]:

#atipicos_ord['2017-01'].head(100)

In [ ]:

#
								# Datos
								#
								# Tengo que usar pandas. rowling
								

In [ ]:

In [36]:

#
								# BORRO DATOS ATIPICOS FALSOS 
								
								#foto = datos.loc['2017-01-27 09:30:00':'2017-01-27 09:30:00'].index
								#datos = datos.drop(foto)
								
								# BORRAMOS ANIO 01/2017
								indice = datos.loc['2017-01', 'MVA'].idxmax()
								datos = datos.drop(indice)
								
								#
								# BORRAMOS DATOS ANIO 10-11-12/2020
								#
								desde = '2020-10-06 17:15:00'
								hasta = '2020-12-14 14:45:00'
								indice = datos.loc[desde:hasta].index
								datos  = datos.drop(indice)
								
								
								desde = pd.to_datetime(desde)
								hasta = pd.to_datetime(hasta)
								duracion = hasta - desde
								print('Se borro: ', duracion, 'que equivale a tantas mediciones tomadas de a 15 min') 
								
								# BORRAMOS ANIO 05/2021
								indice = datos.loc['2021-05', 'MVA'].idxmax()
								datos = datos.drop(indice)
								

Se borro:  68 days 21:30:00 que equivale a tantas mediciones tomadas de a 15 min

In [37]:

# Gráfico boxplot para estacionalidad anual 
								# Cada Box Plot grafica el mes para los 12 anios 
								# Por ejemplo Octubre '10' son todos los valores de Octubre en los 12 anios
								# ==============================================================================
								fig, ax = plt.subplots(figsize=(10, 5))
								plt.style.use('fivethirtyeight')
								
								datos['mes'] = datos.index.month
								datos.boxplot(column='MVA', by='mes', ax=ax)
								datos.groupby('mes')['MVA'].median().plot(style='o-', linewidth=0.8, ax=ax)
								
								# ESTETICA DEL GRAFICO
								plt.xticks(fontsize=12)  #plt.xticks(fontsize=15, rotation=90)
								plt.yticks(fontsize=12)
								#ax.set_ylabel('Demandas')
								ax.set_title('Distribución Demandas por mes 2010 - 2021')
								ax.legend()
								fig.suptitle('');
								

In [23]:

# Gráfico boxplot para estacionalidad anual 
								# Blox Plot para todos los anios
								# ==============================================================================
								
								def Grafica_Caja(pi, pf):
									
									plt.style.use('fivethirtyeight')
									fig = plt.figure()
									fig, ax = plt.subplots(figsize=(30, 5), dpi=100)  # (15, 5)
									
									# PREPARO LOS DATOS
									datos['mes']  = datos.index.month
									datos['anio'] = datos.index.year
									datos2 = datos[pi:pf]
									
									# GRAFICO EL BOXPLOT
									datos2.boxplot(column='MVA', by=['anio','mes'], ax=ax)
									
									# GRAFICO LA LINEA DE MEDIANA
									y = datos2.groupby(by=['anio','mes'])['MVA'].median()
									pos_sup = y.index.shape[0] + 1
									x = np.arange(1, pos_sup)
									ax.plot(x, y, lw=1.5, label='Mediana [MVA]', marker='o')
									#datos3.groupby(by=['anio','mes'])['MVA'].median().plot(style='o-', linewidth=0.8, ax=ax)
									
									# ESTETICA DEL GRAFICO
									plt.xticks(fontsize=12, rotation=90)  #plt.xticks(fontsize=15, rotation=90)
									plt.yticks(fontsize=12)
									#ax.set_ylabel('Demandas')
									ax.set_title('Distribución Demandas por Mes ' +pi+' - '+pf)
									ax.legend() 
									fig.suptitle('');
								

In [24]:

# GRAFICA MAXIMOS SEMANALES
								
								def Grafica_xSemana(pi, pf):
									z = datos.loc[pi:pf, 'MVA'].resample('W').max() # Datos
									x = z.index # Tiempo
								
									plt.style.use('fivethirtyeight')
									fig = plt.figure()
									fig, grafico = plt.subplots(figsize=(30, 5), dpi=100)
									
									grafico.plot(x, z, lw=1, label='Potencia Aparente', marker='o')
									grafico.set_xlabel('[ Meses ]')  # Add an x-label to the axes.
									grafico.set_ylabel('[ MVA ]')  # Add a y-label to the axes.
									grafico.set_title("POTENCIA APARENTE SEMANAL " + pi + ' ' + pf)  # Add a title to the axes.
									#grafico.legend()  # Add a legend.
									#grafico.annotate('Pico Max', xy=(01-31 14:45, 75), xytext=(01-31 15:00, 75), arrowprops=dict(facecolor='black', shrink=0.05)
									#grafico.grid()
									#fig.savefig("Todos_Anios_xSemana.png")
								

In [40]:

# Grafica anual de la Potencia Aparente cada 15 minutos
								
								def Grafica_xmin(pi, pf):
									y = datos.loc[pi:pf, 'MVA']
									x = y.index
									
									plt.style.use('fivethirtyeight')
									fig = plt.figure()
									fig, grafico = plt.subplots(figsize=(30, 5), dpi=100)
									
									grafico.plot(x, y, lw=1.5, marker='o', label='Potencia Aparente x Minuto')
									# grafico.plot(x, y, lw=1, label='Potencia Aparente', marker='o')
									grafico.set_xlabel('[ Meses ]')  # Add an x-label to the axes.
									grafico.set_ylabel('[ MVA ]')  # Add a y-label to the axes.
									grafico.set_title("POTENCIA APARENTE [cada 15 minutos] " + anio)  # Add a title to the axes.
									#grafico.legend(loc='lower left')  # Add a legend.
									
									#grafico.grid(visible='True', which='major', axis='both', color='black', alpha=1, linewidth=0.5)
									grafico.grid(visible='True', which='major', axis='both', color='black', linewidth=0.5)
									grafico.grid(visible='True', which='minor', axis='both', color='black', linewidth=0.5)
									
									x2 = pd.date_range(start=pi+'-01-01', end=pf+'-12-31', periods=5)
									grafico.set_xticks(x2)
									#grafico.annotate('Pico Max', xy=(01-31 14:45, 75), xytext=(01-31 15:00, 75), arrowprops=dict(facecolor='black', shrink=0.05)
									#grafico.grid()
									#fig.savefig("Serie_Anual_15min_"+anio+".png")
								

In [26]:

# Grafica del Pico Maximo Potencia Aparente x DIA
								from matplotlib import ticker
								
								def Grafica_xDia(pi, pf):
									y = datos.loc[pi:pf, 'MVA'].resample('D').max()
									x = y.index
									
									plt.style.use('fivethirtyeight')
									fig = plt.figure()
									fig, grafico = plt.subplots(figsize=(30, 5), dpi=100)
									
									grafico.plot(x, y, lw=1.5, label='Potencia Aparente x Dia', marker='o')
									grafico.set_xlabel('[ Meses ]')  # Add an x-label to the axes.
									grafico.set_ylabel('[ MVA ]')  # Add a y-label to the axes.
									grafico.set_title("POTENCIA APARENTE DIARIA " + anio)  # Add a title to the axes.
									#grafico.legend(loc='lower left')  # Add a legend.
									
									#grafico.grid(visible='True', which='major', axis='both', color='black', alpha=1, linewidth=0.5)
									grafico.grid(visible='True', which='major', axis='both', color='black', linewidth=0.5)
									grafico.grid(visible='True', which='minor', axis='both', color='black', linewidth=0.5)
									
									x2 = pd.date_range(start=pi+'-01-01', end=pf+'-12-31', periods=5)
									grafico.set_xticks(x2)
									
									#grafico.annotate('Pico Max', xy=(01-31 14:45, 75), xytext=(01-31 15:00, 75), arrowprops=dict(facecolor='black', shrink=0.05)
									#grafico.grid()
									#fig.savefig("Serie_Anual_xDia_"+anio+".png")         
								

In [28]:

# GRAFICA DEL PICO MAXIMO POTENCIA APARENTE (CADA 15 MINUTOS) VENTANA 12 hs ANTES y 12 hs DESPUES
								#
								def Grafica_Pico_Anual(pi, pf):
									indice_max = datos.loc[pi:pf, 'MVA'].idxmax()
									delta = pd.Timedelta('12 hours')
									indice_mas12   = indice_max + delta
									indice_menos12 = indice_max - delta
									maximo = datos.loc[indice_menos12:indice_mas12, 'MVA'].max()
									
									y = datos.loc[indice_menos12 : indice_mas12, 'MVA'] # Datos
									x = y.index  # Tiempo
									
									plt.style.use('fivethirtyeight')
									fig = plt.figure()
									fig, ax = plt.subplots(figsize=(10, 3), dpi=100) #fig, ax = plt.subplots(figsize=(10, 5), dpi=100)
									#Rotulo = 'Pico: ' + str(indice_max) + ' ' + str(maximo)
									
									ax.plot(x, y, lw=1, marker='o') #ax.plot(x, y, lw=1, label=Rotulo, marker='o')
									ax.minorticks_on()
									#ax.grid(b=True, which='major', color='blue', alpha=1, linewidth=1)
									#ax.grid(b=True, which='minor', color='beige', alpha=1, linewidth=1)
									# Major Tick Label
									# Minor Tick Label
									plt.xticks(fontsize=10, rotation=90)
									plt.yticks(fontsize=10)
									ax.set_xlabel('[ Horas ]', fontsize=10)  # Add an x-label to the axes.
									ax.set_ylabel('[ MVA ]', fontsize=10)  # Add a y-label to the axes.
									ax.set_title("PICO MAXIMO POTENCIA APARENTE " + str(maximo) + ' ' + str(indice_max), fontsize=12)  # Add a title to the axes.
									#ax.legend(loc='lower left', fontsize=10)  # Add a legend.
									#fig.savefig("Pico_Anual_"+anio+".png")
										
									
									#print(indice_menos12)
									print(indice_max, datos.loc[indice_menos12:indice_mas12, 'MW'].max(), datos.loc[indice_menos12:indice_mas12, 'MVA'].max())
									#print(indice_mas12)
									#maximo = datos.loc[indice_max:indice_max, 'MVA']
									#print(maximo[0])
								

In [41]:

Grafica_xmin('2014-06-18 12', '2014-06-23 16') 
								

								---------------------------------------------------------------------------
								NameError                                 Traceback (most recent call last)
								<ipython-input-41-a46526f9db4c> in <module>
								----> 1 Grafica_xmin('2014-06-18 12', '2014-06-23 16')
								
								<ipython-input-40-57faf5f979cf> in Grafica_xmin(pi, pf)
								     13     grafico.set_xlabel('[ Meses ]')  # Add an x-label to the axes.
								     14     grafico.set_ylabel('[ MVA ]')  # Add a y-label to the axes.
								---> 15     grafico.set_title("POTENCIA APARENTE [cada 15 minutos] " + anio)  # Add a title to the axes.
								     16     #grafico.legend(loc='lower left')  # Add a legend.
								     17 
								
								NameError: name 'anio' is not defined

<Figure size 432x288 with 0 Axes>

In [42]:

Grafica_xmin('2014-06-18', '2014-06-18') 
								

								---------------------------------------------------------------------------
								NameError                                 Traceback (most recent call last)
								<ipython-input-42-1db7e8722eea> in <module>
								----> 1 Grafica_xmin('2014-06-18', '2014-06-18')
								
								<ipython-input-40-57faf5f979cf> in Grafica_xmin(pi, pf)
								     13     grafico.set_xlabel('[ Meses ]')  # Add an x-label to the axes.
								     14     grafico.set_ylabel('[ MVA ]')  # Add a y-label to the axes.
								---> 15     grafico.set_title("POTENCIA APARENTE [cada 15 minutos] " + anio)  # Add a title to the axes.
								     16     #grafico.legend(loc='lower left')  # Add a legend.
								     17 
								
								NameError: name 'anio' is not defined

<Figure size 432x288 with 0 Axes>

In [ ]:

In [128]:

# LANZO LAS GRAFICAS 
								
								anios = ['2010', '2011', '2012', '2013', '2014', '2015', '2016', '2017', '2018', '2019', '2020', '2021']
								
								Grafica_xSemana(anios[0], anios[-1])
								Grafica_Caja(anios[0], anios[-1])
								
								for anio in anios:
									Grafica_xmin(anio, anio)
									Grafica_Caja(anio, anio)
									Grafica_xDia(anio, anio)
									Grafica_Pico_Anual(anio, anio)
								

2010-01-05 14:00:00 36.24 37.32
								2011-12-20 13:30:00 39.13 40.2
								2012-02-16 14:00:00 42.13 43.18
								2013-12-26 15:00:00 41.97 42.7
								2014-01-23 14:00:00 48.77 49.73
								2015-03-11 14:00:00 44.64 46.34
								2016-02-12 13:15:00 45.98 46.79
								2017-03-02 13:15:00 45.16 45.48
								2018-02-07 13:45:00 46.22 46.35
								2019-02-22 14:15:00 45.63 45.77
								2020-12-28 13:15:00 41.71 41.74
								2021-12-29 14:15:00 48.65 48.67

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

In [ ]:

In [129]:

# 
								# GENERAR TABLA EXCEL CON LOS DATOS DEPURADOS
								#
								anios_hist = ['2010', '2011', '2012', '2013', '2014', '2015', '2016', '2017', '2018', '2019', '2020', '2021']
								
								Trafos = pd.Series([25.0, 25.0, 25.0], index=['TD1', 'TD2', 'TD3'])
								POT_INSTALADA = Trafos.sum()
								
								Historica = pd.DataFrame(columns = ['FECHA_HORA', 'P_MW', 'Q_MVAR', 'S_MVA'])
								
								for anio in anios_hist:
									#Anio = pd.to_numeric(anio)
									fecha_hora_i = datos.loc[anio:anio, 'MVA'].idxmax()
									maximo_p     = datos.loc[fecha_hora_i, 'MW'].max()
									maximo_q     = datos.loc[fecha_hora_i, 'MVAR'].max()
									maximo_s     = datos.loc[fecha_hora_i, 'MVA'].max()
									Historica.loc[anio] = [fecha_hora_i, maximo_p, maximo_q, maximo_s]  
									
								# Columnas derivadas
								Historica['POT_INSTALADA'] = POT_INSTALADA
								Historica['CRECIMIENTO_%'] = Historica['S_MVA'].pct_change().mul(100)
								Historica['N-1'] = Historica['POT_INSTALADA'] - Trafos.max()   
								Historica['CARGA_%'] = Historica['S_MVA'] / Historica['POT_INSTALADA'] * 100
								Historica['ANIO'] = pd.to_numeric(Historica.index)
								
								nombres_columnas = ['ANIO', 'FECHA_HORA', 'P_MW', 'Q_MVAR', 'S_MVA', 'CRECIMIENTO_%', 'POT_INSTALADA', 'N-1', 'CARGA_%']
								Historica = Historica.reindex(columns=nombres_columnas)
								Historica
								

Out[129]:

	ANIO	FECHA_HORA	P_MW	Q_MVAR	S_MVA	CRECIMIENTO_%	POT_INSTALADA	N-1	CARGA_%
2010	2010	2010-01-05 14:00:00	36.24	8.89	37.32	NaN	75.00	50.00	49.76
2011	2011	2011-12-20 13:30:00	39.13	9.23	40.20	7.72	75.00	50.00	53.60
2012	2012	2012-02-16 14:00:00	42.13	9.47	43.18	7.41	75.00	50.00	57.57
2013	2013	2013-12-26 15:00:00	41.97	7.88	42.70	-1.11	75.00	50.00	56.93
2014	2014	2014-01-23 14:00:00	48.67	10.21	49.73	16.46	75.00	50.00	66.31
2015	2015	2015-03-11 14:00:00	44.64	12.43	46.34	-6.82	75.00	50.00	61.79
2016	2016	2016-02-12 13:15:00	45.93	8.94	46.79	0.97	75.00	50.00	62.39
2017	2017	2017-03-02 13:15:00	45.16	5.41	45.48	-2.80	75.00	50.00	60.64
2018	2018	2018-02-07 13:45:00	46.22	3.40	46.35	1.91	75.00	50.00	61.80
2019	2019	2019-02-22 14:15:00	45.63	3.57	45.77	-1.25	75.00	50.00	61.03
2020	2020	2020-12-28 13:15:00	41.71	1.59	41.74	-8.80	75.00	50.00	55.65
2021	2021	2021-12-29 14:15:00	48.65	1.51	48.67	16.60	75.00	50.00	64.89

In [130]:

Historica.index
								

Out[130]:

Index(['2010', '2011', '2012', '2013', '2014', '2015', '2016', '2017', '2018',
									   '2019', '2020', '2021'],
									  dtype='object')

In [131]:

pd.set_option('display.float_format', '{:.8f}'.format)
								import statsmodels.formula.api as smf
								lm = smf.ols(formula="S_MVA~ANIO", data = Historica).fit()
								lm.summary2()
								

Out[131]:

Model:	OLS	Adj. R-squared:	0.259
Dependent Variable:	S_MVA	AIC:	62.9670
Date:	2022-06-14 13:43	BIC:	63.9368
No. Observations:	12	Log-Likelihood:	-29.483
Df Model:	1	F-statistic:	4.842
Df Residuals:	10	Prob (F-statistic):	0.0524
R-squared:	0.326	Scale:	9.5674

	Coef.	Std.Err.	t	P>\|t\|	[0.025	0.975]
Intercept	-1102.6916	521.3294	-2.1152	0.0605	-2264.2860	58.9027
ANIO	0.5692	0.2587	2.2006	0.0524	-0.0071	1.1455

Omnibus:	1.100	Durbin-Watson:	1.424
Prob(Omnibus):	0.577	Jarque-Bera (JB):	0.040
Skew:	0.056	Prob(JB):	0.980
Kurtosis:	3.260	Condition No.:	1176765

In [132]:

pd.set_option('display.float_format', '{:.2f}'.format)
								
								Historica['S_MVA_PRED'] = lm.predict(Historica['ANIO'])
								
								nombres_columnas = ['ANIO', 'FECHA_HORA', 'P_MW', 'Q_MVAR', 'S_MVA', 'CRECIMIENTO_%', 'S_MVA_PRED', 'POT_INSTALADA', 'N-1', 'CARGA_%']
								Historica = Historica.reindex(columns=nombres_columnas)
								Historica
								

Out[132]:

	ANIO	FECHA_HORA	P_MW	Q_MVAR	S_MVA	CRECIMIENTO_%	S_MVA_PRED	POT_INSTALADA	N-1	CARGA_%
2010	2010	2010-01-05 14:00:00	36.24	8.89	37.32	NaN	41.39	75.00	50.00	49.76
2011	2011	2011-12-20 13:30:00	39.13	9.23	40.20	7.72	41.96	75.00	50.00	53.60
2012	2012	2012-02-16 14:00:00	42.13	9.47	43.18	7.41	42.53	75.00	50.00	57.57
2013	2013	2013-12-26 15:00:00	41.97	7.88	42.70	-1.11	43.10	75.00	50.00	56.93
2014	2014	2014-01-23 14:00:00	48.67	10.21	49.73	16.46	43.67	75.00	50.00	66.31
2015	2015	2015-03-11 14:00:00	44.64	12.43	46.34	-6.82	44.24	75.00	50.00	61.79
2016	2016	2016-02-12 13:15:00	45.93	8.94	46.79	0.97	44.81	75.00	50.00	62.39
2017	2017	2017-03-02 13:15:00	45.16	5.41	45.48	-2.80	45.38	75.00	50.00	60.64
2018	2018	2018-02-07 13:45:00	46.22	3.40	46.35	1.91	45.95	75.00	50.00	61.80
2019	2019	2019-02-22 14:15:00	45.63	3.57	45.77	-1.25	46.51	75.00	50.00	61.03
2020	2020	2020-12-28 13:15:00	41.71	1.59	41.74	-8.80	47.08	75.00	50.00	55.65
2021	2021	2021-12-29 14:15:00	48.65	1.51	48.67	16.60	47.65	75.00	50.00	64.89

In [133]:

print('Parametros : ', lm.params) 
								print('R2 : ', lm.rsquared)
								

Parametros :  Intercept   -1102.69
								ANIO        0.57    
								dtype: float64
								R2 :  0.3262571112710426

In [134]:

#
								#  AGREGO LOS AÑOS DE PRONOSTICO PROPIAMENTE DICHO
								#
								#Pronostico  = Tablaxls.copy()
								pd.set_option('display.float_format', '{:.2f}'.format)
								anios_pro = ['2022', '2023', '2024', '2025', '2026']
								POT_INSTALADA = Trafos.sum()
								Precaucion    = POT_INSTALADA - Trafos.max()
								
								Anio = pd.Series(pd.to_numeric(anios_pro), index=anios_pro)
								Fecha_Hora = pd.Series(np.nan, index= anios_pro)
								P_MW   = pd.Series(np.nan, index=anios_pro)
								Q_MVAR = pd.Series(np.nan, index=anios_pro)
								S_MVA  = pd.Series(np.nan, index=anios_pro)
								Crecimiento = pd.Series(np.nan, index=anios_pro)
								
								Pronostico = pd.DataFrame({'ANIO':Anio, 
														   'FECHA_HORA':Fecha_Hora, 
														   'P_MW':P_MW, 
														   'Q_MVAR':Q_MVAR,
														   'S_MVA':S_MVA,
														   'CRECIMIENTO_%':Crecimiento}, index=anios_pro)
								
								# El MODELO DE REGRESION LINEAL          
								Pronostico['S_MVA_PRED'] = lm.predict(Pronostico['ANIO'])  # Pronostico['S_MVA_PRED'] = 0.51372727*Pronostico["ANIO"] - 991.015000
								#
								Pronostico['POT_INSTALADA'] = POT_INSTALADA
								Pronostico['N-1'] = Pronostico['POT_INSTALADA'] - Trafos.max()  
								Pronostico['CARGA_%'] = Pronostico['S_MVA_PRED'] / Pronostico['POT_INSTALADA'] * 100 
								
								# nombres_columnas = ['FECHA_HORA', 'P_MW', 'Q_MVAR', 'S_MVA', 'Crecimiento %', 'Pot_Instalada', 'N-1', '% Carga']
								#Tablaxls = Tablaxls.reindex(columns=nombres_columnas)
								
								Pronostico
								

Out[134]:

	ANIO	FECHA_HORA	P_MW	Q_MVAR	S_MVA	CRECIMIENTO_%	S_MVA_PRED	POT_INSTALADA	N-1	CARGA_%
2022	2022	NaN	NaN	NaN	NaN	NaN	48.22	75.00	50.00	64.30
2023	2023	NaN	NaN	NaN	NaN	NaN	48.79	75.00	50.00	65.06
2024	2024	NaN	NaN	NaN	NaN	NaN	49.36	75.00	50.00	65.81
2025	2025	NaN	NaN	NaN	NaN	NaN	49.93	75.00	50.00	66.57
2026	2026	NaN	NaN	NaN	NaN	NaN	50.50	75.00	50.00	67.33

In [135]:

#
								# CONCATENAR LA TABLA HISTORICA + TABLA PRONOSTICO
								#
								Tablaxls = pd.concat([Historica,Pronostico], axis = 0)
								Tablaxls
								

Out[135]:

	ANIO	FECHA_HORA	P_MW	Q_MVAR	S_MVA	CRECIMIENTO_%	S_MVA_PRED	POT_INSTALADA	N-1	CARGA_%
2010	2010	2010-01-05 14:00:00	36.24	8.89	37.32	NaN	41.39	75.00	50.00	49.76
2011	2011	2011-12-20 13:30:00	39.13	9.23	40.20	7.72	41.96	75.00	50.00	53.60
2012	2012	2012-02-16 14:00:00	42.13	9.47	43.18	7.41	42.53	75.00	50.00	57.57
2013	2013	2013-12-26 15:00:00	41.97	7.88	42.70	-1.11	43.10	75.00	50.00	56.93
2014	2014	2014-01-23 14:00:00	48.67	10.21	49.73	16.46	43.67	75.00	50.00	66.31
2015	2015	2015-03-11 14:00:00	44.64	12.43	46.34	-6.82	44.24	75.00	50.00	61.79
2016	2016	2016-02-12 13:15:00	45.93	8.94	46.79	0.97	44.81	75.00	50.00	62.39
2017	2017	2017-03-02 13:15:00	45.16	5.41	45.48	-2.80	45.38	75.00	50.00	60.64
2018	2018	2018-02-07 13:45:00	46.22	3.40	46.35	1.91	45.95	75.00	50.00	61.80
2019	2019	2019-02-22 14:15:00	45.63	3.57	45.77	-1.25	46.51	75.00	50.00	61.03
2020	2020	2020-12-28 13:15:00	41.71	1.59	41.74	-8.80	47.08	75.00	50.00	55.65
2021	2021	2021-12-29 14:15:00	48.65	1.51	48.67	16.60	47.65	75.00	50.00	64.89
2022	2022	NaN	NaN	NaN	NaN	NaN	48.22	75.00	50.00	64.30
2023	2023	NaN	NaN	NaN	NaN	NaN	48.79	75.00	50.00	65.06
2024	2024	NaN	NaN	NaN	NaN	NaN	49.36	75.00	50.00	65.81
2025	2025	NaN	NaN	NaN	NaN	NaN	49.93	75.00	50.00	66.57
2026	2026	NaN	NaN	NaN	NaN	NaN	50.50	75.00	50.00	67.33

In [160]:

# Grafica anual de la Potencia Aparente cada 15 minutos
								# (Trafos + Generacion)
								
								def Grafica_Hist_Pred():
									plt.style.use('fivethirtyeight')
									
									#y = Tablaxls.loc['2010':'2021', 'S_MVA']
									#x_y = Tablaxls.loc['2010':'2021', 'ANIO']
									
									y = Tablaxls.loc[:, 'S_MVA']
									pred = Tablaxls.loc[:, 'S_MVA_PRED']
									p_inst = Tablaxls.loc[:, 'POT_INSTALADA']
									precau = Tablaxls.loc[:, 'N-1']
									
									x = Tablaxls.loc[:, 'ANIO']
									
									fig = plt.figure()
									fig, grafico = plt.subplots(figsize=(15, 5), dpi=100) #fig, grafico = plt.subplots(figsize=(15, 5), dpi=100) 
									#grafico.plot(x, y, lw=1, label='Potencia Aparente', marker='o')
									grafico.plot(x, y, lw=3, label='Pot. Historica', marker='o', color='k')
									grafico.plot(x, pred, lw=3, label='Tend. Lineal', marker='o', linestyle='dotted', color='b')
									grafico.plot(x, p_inst, lw=2, label='Pot.Instalada', color='r')
									grafico.plot(x, precau, lw=2, label='Precaucion N-1', color='y')
									
									grafico.set_ylim ([20,90])
									
									grafico.set_xlabel('[ Años ]')  # Add an x-label to the axes.
									grafico.set_ylabel('[ MVA ]')  # Add a y-label to the axes.
									grafico.set_title("ET San Francisco: Demanda Historica y Pronostico al 2026 ")  # Add a title to the axes.
									grafico.legend(loc='lower right')  # Add a legend.
									#grafico.annotate('Pico Max', xy=(01-31 14:45, 75), xytext=(01-31 15:00, 75), arrowprops=dict(facecolor='black', shrink=0.05)
									#grafico.grid() 
									
									#fig.savefig("Serie_Anual_15min_"+anio+".png")
								

In [161]:

Grafica_Hist_Pred()

<Figure size 640x480 with 0 Axes>

In [219]:

# Grafica anual de la Potencia Aparente cada 15 minutos
								# (Trafos + Generacion)
								
								def Grafica_Hist_Pred2():
									plt.style.use('fivethirtyeight')
									
									#y = Tablaxls.loc['2010':'2021', 'S_MVA']
									#x_y = Tablaxls.loc['2010':'2021', 'ANIO']
									
									y = Tablaxls.loc[:, 'S_MVA']
									pred = Tablaxls.loc[:, 'S_MVA_PRED']
									p_inst = Tablaxls.loc[:, 'POT_INSTALADA']
									precau = Tablaxls.loc[:, 'N-1']
									
									x = Tablaxls.loc[:, 'ANIO']
									
									#fig = plt.figure()
									#fig, grafico = plt.subplots(figsize=(15, 5), dpi=100)
									fig, grafico = plt.subplots(figsize=(7, 4))
								
									
									#grafico.plot(x, y, lw=1, label='Potencia Aparente', marker='o')
									grafico.plot(x, y, lw=2, label='Pot. Historica', marker='o', color='k')
									grafico.plot(x, pred, lw=2, label='Tend. Lineal', marker='o', linestyle='dotted', color='b')
									#grafico.plot(x, p_inst, lw=2, label='Pot.Instalada', color='r')
									grafico.plot(x, precau, lw=2, label='Precaucion N-1', color='y')
									
									#grafico.set_ylim ([20,90])
									
									ax.minorticks_on()
									ax.grid(b=True, which='major', color='blue', alpha=1, linewidth=1)
									ax.grid(b=True, which='minor', color='beige', alpha=1, linewidth=1)
									# Major Tick Label
									# Minor Tick Label
									#plt.xticks(fontsize=10)
									#x_rotulos=[2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022, 2023, 2024, 2025, 2026]
									x_rotulos = np.arange(2010,2027,1)
									plt.xticks(ticks=x, labels=x_rotulos, fontsize='x-small') 
									plt.yticks(fontsize=10)
									
									
									grafico.set_xlabel('[ Años ]', fontsize=10)  # Add an x-label to the axes.
									grafico.set_ylabel('[ MVA ]', fontsize=10)  # Add a y-label to the axes.
									grafico.set_title("ET San Francisco: Demanda Historica y Pronostico al 2026 ", fontsize=12)  # Add a title to the axes.
									grafico.legend(loc='lower right', fontsize=10)  # Add a legend.
									#grafico.annotate('Pico Max', xy=(01-31 14:45, 75), xytext=(01-31 15:00, 75), arrowprops=dict(facecolor='black', shrink=0.05)
									#grafico.grid() 
									
									#fig.savefig("Serie_Anual_15min_"+anio+".png")
								

In [220]:

Grafica_Hist_Pred2()

In [215]:

# Grafica anual de la Potencia Aparente cada 15 minutos
								# (Trafos + Generacion)
								
								def Grafica_Hist_Pred_Seaborn():
									plt.style.use('fivethirtyeight')
									
									#y = Tablaxls.loc['2010':'2021', 'S_MVA']
									#x_y = Tablaxls.loc['2010':'2021', 'ANIO']
									
									y = Tablaxls.loc[:, 'S_MVA']
									pred = Tablaxls.loc[:, 'S_MVA_PRED']
									p_inst = Tablaxls.loc[:, 'POT_INSTALADA']
									precau = Tablaxls.loc[:, 'N-1']
									
									x = Tablaxls.loc[:, 'ANIO']
									
									#fig = plt.figure()
									#fig, grafico = plt.subplots(figsize=(15, 5), dpi=100)
									fig, grafico = plt.subplots(figsize=(7, 4))
									
									sns.lineplot(x=x, y=y, err_style="bars", ci=68)
								

In [216]:

Grafica_Hist_Pred_Seaborn()

In [176]:

Tablaxls.info()
								

<class 'pandas.core.frame.DataFrame'>
								Index: 17 entries, 2010 to 2026
								Data columns (total 10 columns):
								 #   Column         Non-Null Count  Dtype  
								---  ------         --------------  -----  
								 0   ANIO           17 non-null     int64  
								 1   FECHA_HORA     12 non-null     object 
								 2   P_MW           12 non-null     float64
								 3   Q_MVAR         12 non-null     float64
								 4   S_MVA          12 non-null     float64
								 5   CRECIMIENTO_%  11 non-null     float64
								 6   S_MVA_PRED     17 non-null     float64
								 7   POT_INSTALADA  17 non-null     float64
								 8   N-1            17 non-null     float64
								 9   CARGA_%        17 non-null     float64
								dtypes: float64(8), int64(1), object(1)
								memory usage: 2.0+ KB

In [35]:

Tablaxls["RSE"] = (Tablaxls["S_MVA"]-Tablaxls["S_MVA_PRED"])**2
								Tablaxls
								

Out[35]:

	ANIO	FECHA_HORA	P_MW	Q_MVAR	S_MVA	CRECIMIENTO_%	S_MVA_PRED	POT_INSTALADA	N-1	CARGA_%	RSE
2010	2010	2010-01-05 14:00:00	36.24	8.89	37.32	NaN	40.82	75.00	50.00	49.76	12.22
2011	2011	2011-12-20 13:30:00	39.13	9.23	40.20	7.72	42.11	75.00	50.00	53.60	3.64
2012	2012	2012-02-16 14:00:00	42.13	9.47	43.18	7.41	43.40	75.00	50.00	57.57	0.05
2013	2013	2013-12-26 15:00:00	41.97	7.88	42.70	-1.11	44.69	75.00	50.00	56.93	3.98
2014	2014	2014-01-23 14:00:00	48.67	10.21	49.73	16.46	45.99	75.00	50.00	66.31	14.01
2015	2015	2015-03-11 14:00:00	44.64	12.43	46.34	-6.82	47.28	75.00	50.00	61.79	0.88
2016	2016	2016-02-12 13:15:00	45.93	8.94	46.79	0.97	48.57	75.00	50.00	62.39	3.18
2017	2017	2017-01-27 09:30:00	2.95	42.47	75.78	61.96	49.87	75.00	50.00	101.04	671.54
2018	2018	2018-02-07 13:45:00	46.22	3.40	46.35	-38.84	51.16	75.00	50.00	61.80	23.12
2019	2019	2019-02-22 14:15:00	45.63	3.57	45.77	-1.25	52.45	75.00	50.00	61.03	44.64
2020	2020	2020-12-28 13:15:00	41.71	1.59	41.74	-8.80	53.74	75.00	50.00	55.65	144.10
2021	2021	2021-05-04 11:30:00	26.85	52.78	59.22	41.88	55.04	75.00	50.00	78.96	17.50
2022	2022	NaN	NaN	NaN	NaN	NaN	47.74	75.00	50.00	63.66	NaN
2023	2023	NaN	NaN	NaN	NaN	NaN	48.26	75.00	50.00	64.34	NaN
2024	2024	NaN	NaN	NaN	NaN	NaN	48.77	75.00	50.00	65.03	NaN
2025	2025	NaN	NaN	NaN	NaN	NaN	49.28	75.00	50.00	65.71	NaN
2026	2026	NaN	NaN	NaN	NaN	NaN	49.80	75.00	50.00	66.40	NaN

In [36]:

plt.hist((Tablaxls["S_MVA"]-Tablaxls["S_MVA_PRED"]))
								

Out[36]:

(array([1., 2., 5., 1., 2., 0., 0., 0., 0., 1.]),
								 array([-12.00425408,  -8.21241492,  -4.42057576,  -0.6287366 ,
										  3.16310256,   6.95494172,  10.74678089,  14.53862005,
										 18.33045921,  22.12229837,  25.91413753]),
								 <BarContainer object of 10 artists>)

In [37]:

Tablaxls.info()
								

<class 'pandas.core.frame.DataFrame'>
								Index: 17 entries, 2010 to 2026
								Data columns (total 11 columns):
								 #   Column         Non-Null Count  Dtype  
								---  ------         --------------  -----  
								 0   ANIO           17 non-null     int64  
								 1   FECHA_HORA     12 non-null     object 
								 2   P_MW           12 non-null     float64
								 3   Q_MVAR         12 non-null     float64
								 4   S_MVA          12 non-null     float64
								 5   CRECIMIENTO_%  11 non-null     float64
								 6   S_MVA_PRED     17 non-null     float64
								 7   POT_INSTALADA  17 non-null     float64
								 8   N-1            17 non-null     float64
								 9   CARGA_%        17 non-null     float64
								 10  RSE            12 non-null     float64
								dtypes: float64(9), int64(1), object(1)
								memory usage: 1.6+ KB

In [ ]:

Estudio de la demanda historica y pronostico a 5 anios

Pronostico 2022 - 2026 Maximos Potencia Aparente ET SAN FRANCISCO Fecha: 13/10/2021 12:55¶

Seteo Inicial¶

Lectura de las Demandas en 13 Kv San Francisco¶

Depuracion¶

Duplicados¶

Chequeo Mediciones Faltantes¶

Analizar y Encontrar Errores de Medicion usando graficas¶

Deteccion de Valores Erroneos de Medicion¶

Valores constantes¶

Valores Picos Erroneos¶

Normalidad de los datos y Seleccion de la Distribucion¶

Borrar Mediciones Erroneas¶

Graficas Anuales¶

Box Plot Mensual del total de Datos¶

Blox Plot Mensual¶

Grafica Maximos Semanales¶

Grafica Por Minuto (cada 15 Minutos)¶

Grafica por Dia¶

Grafica del Pico Maximo Anual¶

Pronosticos¶

Tabla Historica¶

Regresion Lineal¶

Tabla de Pronosticos¶

Graficas al 2026¶

Conclusiones¶

Contacto